update model

README.md

@@ -24,10 +24,10 @@ model-index:
     metrics:
        - name: Test WER
          type: wer
-         value: 12.29
+         value: 11.83
        - name: Test CER
          type: cer
-         value: 4.23
+         value: 4.09
 ---
 
 # Wav2Vec2-Large-XLSR-53-Portuguese
@@ -81,11 +81,11 @@ for i, predicted_sentence in enumerate(predicted_sentences):
 
 | Reference  | Prediction |
 | ------------- | ------------- |
-| NEM O RADAR NEM OS OUTROS INSTRUMENTOS DETECTARAM O BOMBARDEIRO STEALTH. | NEM UM VADOR MIL O SOUTOS INSTRUMINTOS DE TECTEAM UM BAMBEDER OSTAR |
-| PEDIR DINHEIRO EMPRESTADO ÀS PESSOAS DA ALDEIA | PEDIAR DINHEIRO EMPRESTADO ÀS PESSOAS DA ALDEIA |
+| NEM O RADAR NEM OS OUTROS INSTRUMENTOS DETECTARAM O BOMBARDEIRO STEALTH. | NENHUM VADA ME OS SALTOWS INSTRUMENTOS DE TECTERÁM UM BOMBADEIRO STER |
+| PEDIR DINHEIRO EMPRESTADO ÀS PESSOAS DA ALDEIA | EDIR DINHEIRO EMPRESTADO ÀS PESSOAS DO ALDEIRA |
 | OITO | OITO |
-| TRANCÁ-LOS | TRANCAUDOS |
-| REALIZAR UMA INVESTIGAÇÃO PARA RESOLVER O PROBLEMA | REALIZARAM A INVESTIGAÇÃO PARA RESOLVER O PROBLEMA |
+| TRANCÁ-LOS | TRAM COULTS |
+| REALIZAR UMA INVESTIGAÇÃO PARA RESOLVER O PROBLEMA | REALIZAR UMA INVESTIGAÇÃO PARA RESOLVER O PROBLEMA |
 
 ## Evaluation
 
@@ -156,7 +156,7 @@ My model may report better scores than others because of some specificity of my
 
 | Model | WER | CER |
 | ------------- | ------------- | ------------- |
-| jonatasgrosman/wav2vec2-large-xlsr-53-portuguese | **12.29%** | **4.23%** |
+| jonatasgrosman/wav2vec2-large-xlsr-53-portuguese | **11.83%** | **4.09%** |
 | joorock12/wav2vec2-large-xlsr-portuguese-a | 15.52% | 5.12% |
 | joorock12/wav2vec2-large-xlsr-portuguese | 15.95% | 5.31% |
 | gchhablani/wav2vec2-large-xlsr-pt | 17.64% | 6.04% |

config.json

@@ -43,20 +43,27 @@
   "feat_extract_dropout": 0.0,
   "feat_extract_norm": "layer",
   "feat_proj_dropout": 0.05,
-  "final_dropout": 0.1,
+  "final_dropout": 0.0,
   "gradient_checkpointing": true,
   "hidden_act": "gelu",
   "hidden_dropout": 0.05,
-  "hidden_dropout_prob": 0.1,
   "hidden_size": 1024,
   "initializer_range": 0.02,
   "intermediate_size": 4096,
   "layer_norm_eps": 1e-05,
   "layerdrop": 0.05,
+  "mask_channel_length": 10,
+  "mask_channel_min_space": 1,
+  "mask_channel_other": 0.0,
+  "mask_channel_prob": 0.0,
+  "mask_channel_selection": "static",
   "mask_feature_length": 10,
   "mask_feature_prob": 0.0,
   "mask_time_length": 10,
+  "mask_time_min_space": 1,
+  "mask_time_other": 0.0,
   "mask_time_prob": 0.05,
+  "mask_time_selection": "static",
   "model_type": "wav2vec2",
   "num_attention_heads": 16,
   "num_conv_pos_embedding_groups": 16,
@@ -65,5 +72,5 @@
   "num_hidden_layers": 24,
   "pad_token_id": 0,
   "transformers_version": "4.5.0.dev0",
-  "vocab_size": 47
+  "vocab_size": 46
 }

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a1de92a264a6de3de93f9e65de7e5f490c685f43ebe3e35840b530539e32d790
-size 1262126551
+oid sha256:c8b135f54fc90d03a93c3cee21c655b9652acb2d54619af1520b0cb0d2f6a437
+size 1262122455

vocab.json

@@ -1 +1 @@
-{"<pad>": 0, "<s>": 1, "</s>": 2, "<unk>": 3, "|": 4, "A": 5, "E": 6, "O": 7, "S": 8, "R": 9, "I": 10, "D": 11, "N": 12, "M": 13, "U": 14, "T": 15, "C": 16, "L": 17, "P": 18, "V": 19, "H": 20, "G": 21, "Q": 22, "F": 23, "B": 24, "Ã": 25, "Ç": 26, "Z": 27, "Á": 28, "É": 29, "J": 30, "X": 31, "-": 32, "Í": 33, "Ó": 34, "Ê": 35, "À": 36, "Õ": 37, "Ú": 38, "Ô": 39, "Â": 40, "'": 41, "Y": 42, "Ü": 43, "K": 44, "W": 45, "Ò": 46}
+{"<pad>": 0, "<s>": 1, "</s>": 2, "<unk>": 3, "|": 4, "Ã": 5, "U": 6, "À": 7, "V": 8, "M": 9, "S": 10, "J": 11, "Ü": 12, "C": 13, "'": 14, "L": 15, "P": 16, "W": 17, "Y": 18, "I": 19, "É": 20, "F": 21, "Â": 22, "Z": 23, "D": 24, "H": 25, "R": 26, "K": 27, "Q": 28, "Á": 29, "B": 30, "-": 31, "Ç": 32, "Ê": 33, "E": 34, "G": 35, "T": 36, "O": 37, "N": 38, "X": 39, "Õ": 40, "Ô": 41, "Í": 42, "A": 43, "Ú": 44, "Ó": 45}